mathieu/duct-tape
Watch
1
1
Fork
You've already forked duct-tape
 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Unify model catalogs with Codex, Zen and fzf output #1

Closed
silex wants to merge 0 commits from feat/unified-model-catalogs into main
pull from: feat/unified-model-catalogs
merge into: mathieu:main
Conversation 0 Commits - Files changed -
silex commented 2026-09-19 20:57:08 +00:00
Collaborator
Copy link

Summary

  • Consolidate model scripts into get_models.sh with extensible provider adapters.
  • Support Codex subscriptions, OpenAI, OpenRouter, Anthropic, Google and OpenCode Zen.
  • Add --plain for fzf and --all with provider-prefixed slugs and isolated failures.
  • Preserve default formatting, handle pagination, and remove obsolete entry points.
  • Document credentials, public Zen catalog and aggregate error semantics.

Verification

  • Bash syntax and whitespace checks.
  • Live Codex and Zen catalogs.
  • Mocked provider responses covering formats, pagination, duplicates, credentials, argument conflicts and partial/total failures.

Unrelated local changes to codex-limits and other utilities are excluded.

## Summary - Consolidate model scripts into `get_models.sh` with extensible provider adapters. - Support Codex subscriptions, OpenAI, OpenRouter, Anthropic, Google and OpenCode Zen. - Add `--plain` for fzf and `--all` with provider-prefixed slugs and isolated failures. - Preserve default formatting, handle pagination, and remove obsolete entry points. - Document credentials, public Zen catalog and aggregate error semantics. ## Verification - Bash syntax and whitespace checks. - Live Codex and Zen catalogs. - Mocked provider responses covering formats, pagination, duplicates, credentials, argument conflicts and partial/total failures. Unrelated local changes to codex-limits and other utilities are excluded.
Replace the separate model-fetching scripts with one extensible CLI for
Codex subscriptions, OpenAI, OpenRouter, Anthropic, Google and OpenCode Zen.

- Reuse the native Codex login and app-server catalog without inference
- Fetch Zen models from its documented public endpoint
- Add --plain for fzf while preserving the default list-block format
- Aggregate --all results as provider/model-slug, isolating provider failures
- Sort and deduplicate identifiers and handle Anthropic/Google pagination
- Remove obsolete entry points and document credentials and error semantics

Verified Bash syntax, live Codex and Zen catalogs, and mocked provider,
pagination, formatting, authentication and partial-failure scenarios.
silex closed this pull request 2026-09-19 20:58:50 +00:00

Pull request closed

Please reopen this pull request to perform a merge.
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels   
Agent/Autonomous OK
Safe for an agent to proceed autonomously within repo and risk rules.

  
Agent/Demeter
Demeter can or should work on this.

  
Agent/Hermes
Hermes can or should work on this.

  
Agent/Human
Requires a human decision, credential, review, or action.

  
Agent/Needs Review
Agent output requires review before merge, deploy, or operational action.

  
Area/API
API, backend interface, routes, schemas, or contracts.

  
Area/Auth
Authentication, authorization, identity, permissions, or sessions.

  
Area/CLI
Command-line interface, flags, output formats, or shell workflows.

  
Area/Data
Database, persistence, migrations, data model, import, or export.

  
Area/Docs
Documentation area, docs site, or content structure.

  
Area/Infra
Infrastructure, hosting, networking, system config, or platform operations.

  
Area/UI
Frontend or user interface.

  
CI/Failing
CI currently fails or this work addresses a CI failure.

  
CI/Flaky
Intermittent test or pipeline failure.

  
CI/Green
CI is known passing or validation is complete.

  
CI/Needs Runner
Requires runner setup, capacity, labels, or permissions.

  
CI/Needs Workflow
Requires a CI workflow or pipeline definition.

  
Compat/Backward Compatible
Explicitly backward-compatible change.

  
Compat/Breaking
Breaking change that will not be backward compatible.

  
Compat/Migration
Requires data, config, API, user, or operator migration.

  
Deploy/Ansible
Related to an Ansible-managed deployment path.

  
Deploy/Blocked
Deployment is blocked.

  
Deploy/Homelab
Related to local or homelab deployment.

  
Deploy/Needs Config
Needs configuration, environment variables, service definitions, or operator settings.

  
Deploy/Needs Secret
Needs secret, token, key, credential, or protected variable provisioning.

  
Deploy/Ready
Ready to deploy.

  
Deploy/Rollback
Rollback needed, rollback path involved, or rollback plan required.

  
Kind/Bug
Something is not working. Use for confirmed or suspected defects.

  
Kind/Chore
Maintenance, dependency bumps, cleanup, repository or tool configuration.

  
Kind/Design
Product, UX, API, architecture, or workflow design before implementation.

  
Kind/Documentation
Documentation changes or missing documentation.

  
Kind/Enhancement
Improve existing functionality without changing its core purpose.

  
Kind/Feature
New functionality. Use for net-new product or technical capabilities.

  
Kind/Refactor
Internal restructuring without intended behavior change.

  
Kind/Security
Security issue or security-sensitive change.

  
Kind/Spike
Exploration or prototype intended to reduce uncertainty.

  
Kind/Testing
Tests, QA, validation, fixtures, or coverage work.

  
Ops/Backup
Backup or restore.

  
Ops/Incident
Incident response, outage, emergency fix, or postmortem.

  
Ops/Maintenance
Routine maintenance, housekeeping, or operational cleanup.

  
Ops/Monitoring
Monitoring, alerts, metrics, logs, or observability.

  
Ops/Upgrade
Version upgrade, migration between versions, or upgrade planning.

  
Priority
Critical
 Production-breaking, security-critical, urgent, or otherwise must be addressed immediately.

  
Priority
High
 Important and should be handled soon.

  
Priority
Low
 Low urgency, nice-to-have, or can wait.

  
Priority
Medium
 Normal priority.

  
Priority
Someday
 Worth keeping but not scheduled.

  
Reviewed
Confirmed
 Issue has been confirmed during triage.

  
Reviewed
Duplicate
 This issue or pull request already exists elsewhere.

  
Reviewed
Invalid
 Not valid, not reproducible, or not actionable as filed.

  
Reviewed
Won't Fix
 Deliberately not planned.

  
Risk
Critical
 Very sensitive; requires explicit human approval before action.

  
Risk
High
 Could break production, data, security, compatibility, or important workflows.

  
Risk
Low
 Low operational, data, security, or compatibility risk.

  
Risk
Medium
 Some risk; needs care and appropriate verification.

  
Security/Auth
Authentication or authorization behavior.

  
Security/Disclosure
Sensitive vulnerability or disclosure process.

  
Security/Hardening
Security improvement or hardening.

  
Security/Secret
Secrets, tokens, keys, credentials, or secret handling involved.

  
Size
L
 Large task.

  
Size
M
 Medium task.

  
Size
S
 Small task.

  
Size
XL
 Needs decomposition before implementation.

  
Size
XS
 Tiny change, likely under 30 minutes.

  
Stack/Ansible
Ansible playbooks, roles, inventories, collections, or variables.

  
Stack/Docker
Docker, Compose, images, containers, or registries.

  
Stack/Forgejo
Forgejo/Gitea repositories, API, permissions, webhooks, runners, or settings.

  
Stack/Hermes
Hermes Agent, profiles, gateway, tools, skills, cron jobs, or configuration.

  
Stack/Node
Node.js, npm, frontend tooling, or JavaScript runtime.

  
Stack/Postgres
PostgreSQL database, SQL, indexes, migrations, backups, or tuning.

  
Stack/Python
Python code, packaging, tooling, or runtime.

  
Stack/Systemd
systemd services, timers, units, journals, or service lifecycle.

  
Status
Abandoned
 Work was started but intentionally abandoned or left stale.

  
Status
Blocked
 Something external blocks progress.

  
Status
In Progress
 Work has started.

  
Status
In Review
 Waiting for human, peer, code, or deployment review.

  
Status
Need More Info
 Feedback, reproduction details, acceptance criteria, or a decision is required.

  
Status
Ready
 Ready to be picked up by a human or agent.

No labels
Agent/Autonomous OK
Agent/Demeter
Agent/Hermes
Agent/Human
Agent/Needs Review
Area/API
Area/Auth
Area/CLI
Area/Data
Area/Docs
Area/Infra
Area/UI
CI/Failing
CI/Flaky
CI/Green
CI/Needs Runner
CI/Needs Workflow
Compat/Backward Compatible
Compat/Breaking
Compat/Migration
Deploy/Ansible
Deploy/Blocked
Deploy/Homelab
Deploy/Needs Config
Deploy/Needs Secret
Deploy/Ready
Deploy/Rollback
Kind/Bug
Kind/Chore
Kind/Design
Kind/Documentation
Kind/Enhancement
Kind/Feature
Kind/Refactor
Kind/Security
Kind/Spike
Kind/Testing
Ops/Backup
Ops/Incident
Ops/Maintenance
Ops/Monitoring
Ops/Upgrade
Priority
Critical
Priority
High
Priority
Low
Priority
Medium
Priority
Someday
Reviewed
Confirmed
Reviewed
Duplicate
Reviewed
Invalid
Reviewed
Won't Fix
Risk
Critical
Risk
High
Risk
Low
Risk
Medium
Security/Auth
Security/Disclosure
Security/Hardening
Security/Secret
Size
L
Size
M
Size
S
Size
XL
Size
XS
Stack/Ansible
Stack/Docker
Stack/Forgejo
Stack/Hermes
Stack/Node
Stack/Postgres
Stack/Python
Stack/Systemd
Status
Abandoned
Status
Blocked
Status
In Progress
Status
In Review
Status
Need More Info
Status
Ready
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
mathieu/duct-tape!1
No description provided.